Académique Documents
Professionnel Documents
Culture Documents
S
P
A
C
E
M
a
g
a
z
i
n
e
1
/
2
0
1
3
www.dspace.com
1/2013
MAGAZINE
dSPACE
Siemens Development Process
According to ISO 26262
Mazda More Compression,
Less Fuel
Thales Alenia Space
Computers in Orbit
Get a Grip on Your Data
with dSPACE SYNECT
II
1401/1511/1512 with an inserted
DS1552 Multi-I/O Module can now
be used for FPGA-based cylinder
pressure measurement and indicati-
on (MABXII Cylinder Pressure Indica-
tion Solution). The solution supports
angle-synchronous pressure measu-
rement with 0.1 max. resolution.
The evaluation of the pressure va-
lues can be controlled by several
algorithms in real-time, such as
released heat, mass fraction burned,
the indicated mean effective pres-
sure, etc.
SCALEXIO Solutions for Aero-
space Applications
dSPACE has added the rst I/O inter-
faces for aerospace applications to
its SCALEXIO
hardware-in-the-loop
(HIL) system. The new interfaces
provide access to ARINC-429 and
MIL-STD-1553 networks, and are
based on tried-and-trusted PMC/
XMC modules from Avionics Inter-
face Technologies (AIT). These mod-
ules use the PCI(e) interfaces in the
SCALEXIO Processing Unit to en-
sure the optimum bandwidth for
data exchange with the real-time
model. To guarantee real-time capa-
bility, dSPACE developed dedicated
QNX drivers for this application
case.
Please let us know your opinion on the quality of the
dSPACE Magazine. Just ll out the enclosed reply card
and return it to us. You can also use the card to order
any other information by post.
Thank you!
You can also give us your feedback online. For further
information, visit:
www.dspace.com/magazine
Release information on dSPACE products is available at:
www.dspace.com/releases
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
At the end of 2011, we announced
our goal of doing something to
more effectively handle the ood of
data objects that arises in every part
of the model-based design process.
For developers to keep track of every-
thing, all this data has to be managed
and linked from the requirements to
the rst design, then to the simula-
tion and parameterization, and nally
to the tests. Access has to be con-
trolled by user privileges, which
requires check-ins and check-outs.
There has to be a secure and reliable
way for a large number of people to
work together, even if they are not
in the same room or even the same
geographical location. To top it all,
variants also need to be accurately
represented and managed, not just
the versions.
Some of these issues have been
solved for software development,
but up until now there was no com-
prehensive solution dedicated to
model-based design as we under-
stand it, including the left-hand side
of the V-cycle.
dSPACE has been aware of this
problem for a long time. Other pri-
orities unfortunately kept us from
tackling it earlier. But since 2011,
we have been giving full attention
to this with SYNECT, our collabora-
tion and data management software.
Because SYNECT builds on a proven
platform, we are able to quickly pro-
duce one solution after another to
cover different topics and problem
areas. All with full knowledge of
what model-based design means
and of the data and semantics that
have to be handled.
The rst modules, for test manage-
ment and variant management,
were released last year. A solution
for signal and parameter manage-
ment followed a short while ago.
Support for model management will
be the next to join them. And the
rst public customer feedback is
already coming in. Attendees at our
User Conference in Stuttgart last
January were shown how our cus-
tomer ABB uses SYNECT Test Man-
agement to test their controls for
locomotive drives. ABB manages the
test plans in SYNECT, controls the
test execution with AutomationDesk,
collects and manages the test results
in SYNECT, and generates test reports
from the results. Also at the confer-
ence, Audi Electronic Ventures, a
subsidiary of AUDI AG, described
how they perform signal and param-
eter management in SYNECT, even
for variants.
There is strong interest in SYNECT.
So our SYNECT team is busy analyzing
our customers situations, collecting
customer requirements and putting
them straight into the product, and
also supporting pilot projects and
production projects. If data manage-
ment in model-based development
projects is causing you sleepless
nights, you just need to ask us
for help!
Dr. Herbert Hanselmann
President
PAGE
3 EDITORIAL
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE 4 CONTENTS
IMPRINT
dSPACE MAGAZINE is published periodically by:
dSPACE GmbH Rathenaustrae 26
33102 Paderborn Germany
Tel.: +49 5251 1638-0
Fax: +49 5251 16198-0
dspace-magazine@dspace.com
www.dspace.com
Responsible in terms of press law:
Bernd Schfers-Maiwald
Project Manager: Andr Klein
Authors: Thorsten Bdeker, Ralf Lieberwirth,
Sonja Lillwitz, Thomas Phlmann, Julia Reinbach,
Dr. Gerhard Rei
Co-Workers on this issue:
Dirk Berneck, Holger Krisp, Bjrn Mller
Editors and Translators: Robert Bevington,
Stefanie Bock, Dr. Michelle Kloppenburg,
Christine Smith
Design: Krall & Partner, Dsseldorf, Germany
Design and Layout: Sabine Stephan
Print:
Media-Print Group GmbH, Paderborn
Copyright 2013
All rights reserved. Written permission is required
for reproduction of all or parts of this publication.
The source must be stated in any such reproduc-
tion. dSPACE is continually improving its products
and reserves the right to alter the specications of
the products contained within this publication at
any time without notice.
dSPACE is a registered trademark of dSPACE GmbH
in the United States or other countries, or both.
See www.dspace.com/goto?trademarks for a list
of further registered trademarks. Other brand
names or product names are trademarks or regis-
tered trademarks of their respective companies
or organizations.
FORD | PAGE
SIEMENS | PAGE
12
28
20
THALES ALENIA SPACE | PAGE
UNIVERSITY OF STUTTGART | PAGE
SYSTEMS | PAGE
BATTERY MANAGEMENT
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
3 EDITORIAL
by Dr. Herbert Hanselmann,
President
Customers
6 MAZDA
Aiming High
Function optimization and validation for
management of gasoline engines with a
compression ratio of 14:1
12 SIEMENS
Process for Functional Safety
Model-based software development for
electric drivetrains according to ISO 26262
20 THALES ALENIA SPACE
Project Orbit
Autocoding satellite software with
TargetLink
24 FORD
Fuel Cell Revolution
Toolchain for developing a fuel cell
control system
28 UNIVERSITY OF STUTTGART
Architecture with Brains
Control concept for ultra-lightweight
structures
34 UNIVERSITY OF KAISERSLAUTERN
Robots in the Rose Garden
Autonomous vehicles for horticulture and
agriculture
Products
40 BATTERY MANAGEMENT SYSTEMS
Perfect Balance
Controlling Li-ion cell voltages during
prototyping
44 CONTROLDESK
A Clear View
New ControlDesk instruments for auto-
motives, aviation and industry
46 AUTOMOTIVE SIMULATION MODELS (ASM)
Virtual Winter Tests
Test drives in virtual reality with dSPACE
Automotive Simulation Models
50 ETHERNET CONFIGURATION PACKAGE
Easy Networking
New Ethernet Conguration Package
makes network conguration easier
Business
54 SCALEXIO
Success in Action
SCALEXIO in customer projects
56 USER CONFERENCE
Driven by Innovation
OEMs and suppliers present technological
trends at the 7th dSPACE German User
Conference
62 COMPACT NEWS
PAGE
5
Contents
24
FORD | PAGE
28
40
SYSTEMS | PAGE
BATTERY MANAGEMENT
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Aiming
High
MAZDA PAGE 6
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Function optimization and validation for
management of gasoline engines with
a compression ratio of 14:1
Mazda is pursuing two main goals for its automobiles: an enjoyable driving
experience on the one hand, and outstandingly low fuel consumption and
emission levels on the other. By 2015, the company plans to reduce con-
sumption by 30 % compared with 2008. So Mazda launched the SKYACTIV
program, designed to optimize all the components necessary for reaching
these goals. dSPACE Simulators and the Automotive Simulation Models
are playing a major role in this.
PAGE
7
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Figure 2: HIL simulation is a key component of quality assurance. MIL simulation opens new
possibilities for early function verication.
Visions for Combustion Engines
In conventional internal combustion
engines, 70 % to 80 % of the fuels
energy is lost even before it reaches
the wheels. Mazda is taking on this
unpleasant fact and aims to make
the combustion process more ef-
cient by reaching ideal combustion
at the highest possible compression
ratio. By 2015, the efciency of the
Mazda eet should be increased by
30 % compared to the levels in
2008. These goals are dened in
SKYACTIV-G, a research and devel-
opment program dedicated to opti-
mizing gasoline engines (gure 1).
Not only active engine improvement
but also further consumption-reducing
measures such i-stop (idling stop sys-
tem) and i-ELOOP (capacitor-based
energy regeneration system) will all
contribute to increasing efciency.
Optimized Engine Management
A high compression ratio consider-
ably improves thermal efciency.
The compression ratio of current
gasoline engines is around 10:1 to
12:1. If the ratio is raised from 10:1
to 15:1, the thermal efciency the-
oretically increases by about 10 %.
Increased torque loss due to higher
knocking is one of the reasons why
this is difcult to achieve in reality.
Knocking is abnormal combustion
during which the air-fuel mixture
ignites prematurely. It is caused by
the effects of high temperatures
and pressures. Improvements in
combustion engines therefore
require comprehensive engine man-
agement measures. To achieve the
best management possible, the
engine ECU software had to be
completely revised and expanded
to include many new functions.
Improved control strategies were
developed to deal with knock, ping
and possible misres. This required
that the variable valve timing (VVT)
system has an optimal control for
the intake air and exhaust gases.
Optimizations in the Develop-
ment Process
Because the complexity of the func-
tions to be developed is increasing,
the software development process
also needs to meet the new devel-
opment requirements. Mazda there-
fore decided to switch to a fully
model-based development method.
This involves specialized tools,
appropriate processes and enor-
mous engineering know-how. The
benets of this are evident in sev-
eral areas.
Model-Based Function Develop-
ment and Validation
There are established methods for
Figure 1: The MAZDAs SKYACTIV program aims at optimizing elementary vehicle components.
MAZDA PAGE 8
SKYACTIV-G SKYACTIV-DRIVE SKYACTIV-BODY
SKYACTIV-D SKYACTIV-MT SKYACTIV-CHASSIS
ENGINE TRANSMISSION PLATFORM
System design System verication
Component design
Coding and installation
Build in quality Verify quality
Component verication
Validation and calibration
in vehicle and engine
Autocode generation
Model-based calibration
Verication
MIL
HIL
Denition of requirements
Descriptive check and
model management
RT-MIL
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
validating the quality of functions
developed by model-based develop-
ment. In the V-cycle, each develop-
ment stage has a corresponding
test phase. Proven test methods
include procedures such as hard-
ware-in-the-loop (HIL) and model-
in-the-loop (MIL) simulation (gure 2).
Both of these methods were used
for the SKYACTIV program. dSPACE
simulators equipped with Automo-
tive Simulation Models (ASM) were
also incorporated. The testing poten-
tial of simulators goes much further,
though. They can also be utilized
for function optimization. To imple-
ment the objectives of SKYACTIV-G,
clear concepts and approaches were
outlined. For the targeted compres-
sion ratio of 14:1, it was necessary
to improve the cylinder charge level.
The throttle valve and the position
of the VVT valve are the inuencing
factors here. However, there are also
interference factors whose effects
need to be known and accounted
for, such as deposits, component
settles on valves and valve seats and
obstructs the ow of the gases.
They have the greatest impact at a
low valve opening. In previous func-
tion developments without a HIL
simulator, investigations were per-
formed with real components, and
PAGE
9
By using dSPACE systems right from the design phase for controller algo-
rithms, we were able to use the same test scenarios throughout the whole
development process.
Satoshi Komori, Mazda
Figure 3: Values inuencing the charge of a combustion engine.
just preparing the parts and vehicles
was expensive enough. Ever since
HIL simulators have been available
for this, these investigations can be
performed through simulation (g-
ure 4). The plant model used for
this simulates the mass ow
tolerances, and the ambient tem-
perature and pressure (gure 3).
Function Analysis and Optimiza-
tion with HIL
As an interference factor, deposits
are a prime example of how HIL
simulation can be used to investigate
and optimize functions. Deposits are
caused by combustion residue that
Figure 4: Combining physical and statistical
models creates precise simulation models
that are easy to handle.
Physical
model
Inlet
Physical
model
Cylinder
Statistical model
Exhaust
Injected volume
Ignition timing
Inlet and exhaust VVT opening
Etc.
A/F ratio
Gas tem-
perature
Etc.
Inlet
manifold
pressure
Etc.
Example: Inlet air
volume control system
Control factors
Output Input
Error factors
System
Throttle opening
VVT opening
Etc.
Deposit
Variation in components
External air temperature, atmospheric pressure
Etc.
Target charging efciency Actual charging efciency
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Conclusion and
Outlook
For the SKYACTIV program,
Mazda relies on MIL and HIL sim-
ulation with the dSPACE Simulator
and used dSPACE Automotive
Simulation Models (ASM) for
fast, efcient setups. The test
equipment is used for function
optimization and validation.
Overall, Mazda was able to revise
the individual components, such
as the engine ECU and the com-
plex electrics/electronics (E/E) sys-
tem network, and validate them
completely and comprehensively.
The particular benets of HIL
testing are the automated and
reproducible test runs. Initial expe-
rience has shown that early test-
ing and test reusability have
enormous potential for increas-
ing efciency. This method has
proven its worth in vehicles like
CX-5 and Mazda 6, which all ben-
et from superb SKYACTIVE-G
gasoline engines with a com-
pression ratio of 14:1. The test-
ing processes and systems will be
optimized and expanded for fu-
ture developments. It is already
clear that they are essential for
the further development of E/E
systems and will be given addi-
tional roles in the development
process.
through the valves and the com-
bustion process in the cylinder with
a high level of accuracy. To account
for the effects of deposits, a new
additional model component was
implemented and integrated into the
plant model. The models open
structure made this very easy. Now
the various investigations can be
performed and appropriate function
optimizations made. If necessary, the
tests can even be automated.
A comparison with the conventional
approach showed that using a HIL
simulator is eight times more ef-
cient. In addition, the HIL simulator
helped to recognize controller arti-
facts, which let the engineers elimi-
nate potential problems. By using
HIL simulation early and consistently,
MAZDA PAGE 10
Figure 5: Test bench for the SKYACTIV-G ECUs.
Figure 6: Stations where real-time MIL simulation is performed.
it was possible to reliably determine
the cylinder charge for the various
operating modes and optimize the
overall controller strategy. Even
though long-term tests were run for
interference factors such as tolerances
and wear, the proactive use of HIL
simulation for function validation in
engine control would bring savings
of 2,500 hours.
Process Optimization with
Real-Time MIL
A further test method demonstrates
the exibility and testing capabilities
that simulation brings. The method
is based on MIL simulation and was
extended to include real-time MIL
simulation. The plant model and the
controller model both run on the
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
simulator, where they form a closed
loop (gure 6). The real-time capa-
bility naturally depends on the pro-
cessing power. In this case, the
processor was a DS1006 Quad-Core
Processor Board. The advantage of
using common dSPACE products is
that test cases can be reused
between HIL and MIL. The method
was used to promptly solve problems
that occurred in the actual vehicle.
The HIL environment, which is
closer to the actual vehicle, is used
to analyze the mechanism of prob-
lems and nd countermeasures.
Then the countermeasures are
introduced to the control models,
where the positive/negative effects
are veried in the MIL environment.
The engineers were able to carry
out a series of verication processes
efciently by reusing the HIL test
cases. In order to reduce problems
in actual vehicles, we are currently
investigating ways to utilize MIL
more effectively in early stages on
the left side of the development
V-cycle.
Satoshi Komori,
Keisuke Yayoi,
Mazda
PAGE
11
SKYACTIV-G posed enormous challenges regarding development time.
Products and services from dSPACE supported the model-based develop-
ment, which enabled us to overcome the hurdles.
Keisuke Yayoi, Mazda
Team at Mazda
Mazda HIL MIL team members who worked on the SKYACTIV control systems.
From left to right: Yoichi Teraoka, Yasuhiro Doi, Keisuke Yayoi, Satoshi Komori,
Takuro Miyoshi.
Satoshi Komori
Satoshi Komori is Assistant Manager
of the Powertrain Control System
Engineering Group at Mazda in
Hiroshima, Japan.
Keisuke Yayoi
Keisuke Yayoi is Senior Staff of the
Powertrain Control System Engineer-
ing Group at Mazda in Hiroshima,
Japan.
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Process
Model-based software development for electric
drivetrains according to ISO 26262
SIEMENS PAGE 12
for Functional Safety
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
The Siemens Drive Technolo-
gies Division has added model-
based development to their
classic software development
process for implementing
safety-critical vehicle func-
tions. With the support of
dSPACE TargetLink Strategic
Partner, Model Engineering
Solutions GmbH, Siemens
Drive Technologies dened
a process on the basis of
dSPACE TargetLink that sat-
ises the requirements of
ISO 26262 (Road Vehicles
Functional Safety).
PAGE
13
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Safety-Critical Software in
a Vehicle
Many suppliers in the automotive
industry are either already using
model-based development for their
safety-critical software, or facing the
challenge of having to add model-
based methods to their software
development processes in the future.
Model-based development brings
numerous proven advantages, such
as higher software quality and easier
software maintenance. With the
introduction of the new ISO 26262,
there is now an international stan-
dard for the functional safety of in-
vehicle electric/electronic systems.
ISO 26262 is derived from the IEC
61508 safety standard, adapted to
meet the specic conditions in a
vehicle. Unlike the IEC standard, ISO
26262 explicitly applies to model-
based development. This new stan-
dard clearly states what the devel-
opment process for safety-critical,
in-vehicle software must achieve to
guarantee the softwares functional
safety with regard to the various
Automotive Safety Integrity Levels
(ASIL, classied in levels from A to D).
However, the standard leaves the
question of how this can be imple-
mented largely unanswered. Practi-
cal experience shows that there is
no universally accepted approach for
applying the standard. Only project-
specic solutions can prove success-
ful, because they also take the com-
panys or the departments existing
processes and tool chains into con-
sideration.
Electrical Vehicle Components
from Siemens
Siemens Drive Technologies devel-
ops, produces and markets key com-
ponents for e-drive vehicles. This is
done from a project-based point of
view, keeping in mind the specic
requirements of the automotive indus-
try customers in each project. Its
portfolio covers a wide range from
electric motors and power electronics
to intelligent onboard charging tech-
nology (gure 1). Siemens has already
established a classic development
process for software functions, pri-
marily oriented around the wide-
spread V-cycle. This development
process complies with the Automo-
tive SPICE guidelines (aSPICE: a ver-
sion of the international standard
ISO/IEC 15504 SPICE, specially
adapted to automotive needs) and
CMMI-Dev (Capability Maturity
Model Integration for Development).
Virtually all vehicle manufacturers
demand compliance with these two
maturity models, so compliance is
absolutely essential.
Requirements for Model-Based
Development of Safety-Critical
Software
Siemens Drive Technologies wished
to integrate model-based develop-
ment into its existing classic develop-
ment process because of the plan to
develop a larger portion of safety-
critical vehicle functions with the
model-based methods of MATLAB
/
Simulink
/Stateow
and dSPACE
TargetLink
in conjunction
with an automatic code generator.
TargetLink Best in Test
Thales Alenia Space investigated
the suitability of several code gen-
erators. These had to fulll several
criteria: As well as supporting the
workow, the code generator had
to be exible with regard to naming
rules, code structure and function
interfaces so that it could be inte-
grated into the existing coding
scenario. Other criteria were the
readability and reliability of the gen-
erated code. The evaluation of sev-
eral code generators revealed that
TargetLink was the best match for
Thales Alenia Spaces evaluation
criteria. TargetLinks rich graphical
user interfaces provide intuitive and
powerful options, including support
for structures, pointers and access
functions. Moreover, the separation
generation in 2004. The goal was
a process where different develop-
ment teams could exchange infor-
mation in a dened, unambiguous
format. At the same time, responsi-
bilities for function development and
software development had to be
coding of satellite software were
becoming obvious (it was difcult
to retrieve information from earlier
projects, to perform software main-
tenance, and so on). Thales Alenia
Space therefore started evaluating
software tools for automatic code
Figure 1: One cause of disruption in satellite electronics is solar ares that hurl charged par-
ticles as far as the Earth (the photo shows a large are that occurred on August 31, 2012).
Figure 2: The clearly structured development
process ensures clean separation of responsi-
bilities, an unambiguous ow of information
and the complete traceability of all the work
steps.
Step 1:
Develop the oating-point Simulink model
Step 2:
Test the model as model-in-the-loop (MIL tests)
F
u
n
c
t
i
o
n
D
e
v
e
l
o
p
e
r
s
MIL
Step 3:
Congure data types, variable names, function names,
function interfaces
Step 4:
Generate the code with TargetLink
S
o
f
t
w
a
r
e
D
e
v
e
l
o
p
e
r
s
P
o
s
s
i
b
l
e
m
o
d
e
l
r
e
w
o
r
k
s
Step 5:
Test the generated code as software-in-the-loop
(SIL tests)
Step 6:
Integrate the nished code on the platform computer
SIL
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE
23 PAGE
23
between Simulink data types for
simulation and TargetLink data
types for code generation allows
round trips between different devel-
opment teams, such as function
developers on one side and soft-
ware developers on the other, in
which both teams can work on
the same model.
Streamlined Development Process
The process begins with the func-
tion developers developing a oat-
ing-point Simulink model, which
they test with model-in-the-loop
(MIL) tests in the second step (g-
ure 2). Then they pass the model to
the software developers, who carry
out the third step conguring the
data types, variable names, function
names and function interfaces. The
fourth step is automatic code gen-
eration with TargetLink, followed by
software-in-the-loop (SIL) testing in
the fth step. If the result of the SIL
test indicates that the model needs
reworking, the software developers
pass it back to the function devel-
opers again. This is where the new
process reveals its great strength,
because although the model has to
go back to the drawing board,
the work already done by the soft-
ware developers is not lost thanks
to the cleanly dened exchange
formats and work steps.
As soon as all the model modica-
tions are completed and the SIL tests
have run successfully, the sixth and
nal step is to integrate the nished
code on the platform computer.
Real-World Success
The rst project with TargetLink
was the development of software
for two infrared Earth observation
satellites (SPIRALE), which were
launched into orbit in 2009 by an
Ariane 5. That same year, Thales
Alenia Space also decided to use
TargetLink for a new project called
Sentinel 3. Sentinel 3 is part of
Global Monitoring for Environment
Christophe Moreno
Christophe Moreno is Chief On Board
SW Architect of the SW Solution Com-
petence Center at Thales Alenia Space
in Cannes, France.
Arnaud Dupuy
Arnaud Dupuy is On Board SW Project
Manager of the Sentinel 3 Platform at
Thales Alenia Space in Cannes, France.
and Security (GMES), a mission of
the European Space Agency (ESA).
Thales Alenia Space has already
generated 12,000 lines of code for
this. Sentinel 3 will be launched
later this year.
TargetLink Firmly Established
Model-based design and automatic
code generation have since become
the established practice at Thales
Alenia Space. Using models that
can be shared by different teams
makes work very efcient. And with
Figure 3: The code generator TargetLink has convincingly demonstrated its advantages
in two projects.
TargetLink, the code structure and
naming rules can be congured
exibly. Thales Alenia Space was
therefore able to integrate the code
into their existing framework with-
out having to modify it. The new
process has already considerably
boosted productivity, and is expected
to produce further major benets
when earlier models are reused in
future projects.
Arnaud Dupuy,
Christophe Moreno
Thales Alenia Space
Thales Alenia Space Projects with TargetLink
Project SPIRALE (two identical twin satellites)
Purpose: Observing the Earth in the infrared range
5,000 lines of code generated by TargetLink
Satellites already launched (2009)
Project SENTINEL 3
Purpose: Measuring the temperature of ocean
and land surfaces, etc., and the topology of ocean
and ice surfaces
12,000 lines of code generated by TargetLink
Launch planned for 2013
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Fuel Cell
For the past 20 years, Ford has gained a wealth of experience from their
research and development of fuel cells for in-vehicle use. Fords researchers
have tested several concepts in the laboratory and on the road. Thanks to
this continuous development work, the new technology is approaching
commercial viability.
Toolchain for Developing
a Fuel Cell Control System
PAGE 24 FORD
Revolution
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Fuel Cell Systems and Their
Control
Fuel cells transform chemical energy
into electrical energy. When reactants
such as hydrogen and oxygen are
supplied to a stack of fuel cells, the
resulting electrons can be harnessed
to propel automobiles. The electronic
control regulates an efcient delivery
of the reactants to the fuel cell by
monitoring and controlling the ow,
concentration and pressure.
Control System Challenges
Early prototype designs like the
P2000 used standard ECUs with pur-
pose-built, customized I/O for fuel
cell applications. But this architecture
left little room for experimentation.
When Ford began internal develop-
ment of control algorithms for fuel
cells, typical constraints such as tim-
ing, personnel, logistics, and budgets
made it necessary to switch from a
standard electronic control unit (ECU)
to a exible rapid prototyping system.
A well-integrated control system was
desired that could drive multiple
valves, which control and route the
hydrogen, air, and water within the
fuel cell system. At the same time,
high exibility and modularity was
absolutely essential because I/O
demands can quickly change at any
time during the research and devel-
opment of electronic systems. Another
challenge was the planned simulta-
neous introduction of a purely
model- based work method.
Ultimate Flexibility with Rapid
Control Prototyping
Ford decided on a rapid control
prototyping (RCP) system consisting
of a dSPACE MicroAutoBox II and
dSPACE RapidPro Power Unit, as
well as Simulink
from MathWorks
(gure 1). The determining factor
behind this decision was the posi-
tive experience with dSPACE RCP
systems in several research and
development divisions at Ford. The
system provides the desired exibil-
ity because it can be adapted to
PAGE
25
The Ford REFL3X concept car, introducing
an electric propulsion system.
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE 26
RapidPro standard modules that are
used in regular combustion engine
applications. For the majority of the
actuators, low-side driver (LSD)
modules were used. The electronic
throttle body in the air delivery
subsystem was controlled by a
full-bridge driver (FBD) module.
Since the fuel cells are driven by two
different supply voltages, the split
voltage bus provided by RapidPro
was especially benecial: two sepa-
rate power rails feed the two respec-
tive supply units (gure 2). These are
just some of the advantages of rapid
prototyping systems.
They are examples of the experience
and ingenuity of dSPACE engineers
who were on hand from project
inception through the detailed imple-
mentation, supporting wherever
and whenever needed.
Convenient Conguration
Operating and conguring the
RapidPro Power Unit for the applica-
tion at hand is done with dSPACEs
software CongurationDesk and its
graphical, intuitive user interface.
For example, CongurationDesk
allows a user to assign channel
names corresponding to actuators
and sensors and to congure set-
tings for each of the channels. At
any time, the user can export a
channel list indicating the location
of modules in their corresponding
slots and the channel names with
their assigned pins. Conguration-
Desk makes handling the congura-
tion and documentation tasks very
easy.
Diagnostics
With its diagnostic functions, the
RapidPro Power Unit uses a serial
peripheral interface (SPI) to prepare
data that can be monitored immedi-
ately in CongurationDesk. What is
more valuable, though, is that this
data can be time-aligned to other
system events and thus logged bet-
ter. For this task, the RTI blocksets
FORD
dSPACE engineers were by our side from project
inception through the detailed implementation,
supporting wherever and whenever needed.
Kurt Osborne, Ford
Figure 2: Two rails in RapidPro allow for twin independent power supplies.
meet the fuel cell control system
requirements by adding any neces-
sary supplementary I/O and inter-
face modules, power units and
signal conditioning units.
In-Project Benets of Using
RapidPro
The signals from the sensors and
actuators of the fuel cell could be
processed directly with the same
RS1 Fuel Cell System
Actuators Sensors
MicroAutoBox II
Break-Out Box
RapidPro Power Unit
Break-Out Box
Figure 1: Interface from the RapidPro Power Unit to MicroAutoBox II.
GND
1
RapidPro Power Unit
U
BAT1
U
BAT2
Front panel
GND
2
Rail 1 Rail 2
1
2
3
4
5
6
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE
27
for Simulink are available to poll the
I/O channels. The recorded data is
processed further in ControlDesk.
To use the diagnostic information
in other test system controllers, it
is retransmitted via CAN messages.
Evaluation and Next Steps
The rapid prototyping systems
MicroAutoBox and RapidPro proved
themselves to be extremely useful
for controlling fuel cell systems dur-
ing research and development at
Ford. All of the specied functions
were able to be implemented in the
fuel cell system, which is attribut-
able, in part, to the coordinated tool
chain. The RapidPro Power Unit ful-
lled all of the needs for a exible
and recongurable power stage.
The RapidPro systems allowed a
small team with a limited budget to
implement and test new fuel cell
concepts without focusing on the
optimization details that would have
been required for ECUs. The switch
to one common tool chain was sur-
prisingly easy.
In near future, the system will undergo
validation testing in freezing ambient
conditions. If additional components
or novel strategies become neces-
sary, the congurability of the Rapid-
Pro Power Unit will allow rapid inte-
gration of those components into
the control system.
One of the next steps could be inte-
grating the fuel cell system into a
demonstration vehicle. In this case,
Ford will be able to use the RapidPro
Power Unit from the very start.
Thanks to its short boot times, the
RapidPro Power Unit is very fast,
making it compatible with vehicle
applications. Furthermore, the Rapid-
Pro Power Unit can be selectively
disabled by the MicroAutoBox thus
saving energy and improving overall
energy consumption.
Kurt Osborne
Dr. Milo Milai
Ford Motor Company
Figure 3: Lab setup of the fuel cell system with MicroAutoBox and RapidPro.
Kurt Osborne
Kurt Osborne is the fuel cell controls
technical expert and is the leader of global
model-based fuel cell system control design
at Ford Motor Company in Dearborn (MI),
USA.
Dr. Milo Milai
Dr. Milo Milai is responsible for fuel
cell control algorithm development and
implementation at Ford Motor Company
in Dearborn (MI), USA.
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
SEITE 28 PAGE 28 UNIVERSITY OF STUTTGART
Lightweight, safe, sustainable, and economical
with materials modern architectural structures
like buildings and bridges have to be all that.
But they also have to stand up to constant stresses
such as snow and wind. Researchers at the Uni-
versity of Stuttgart are working to develop light-
weight load-bearing structures that respond to
external loads.
Figure 2: Active support for the Stuttgart SmartShell, with three hydraulic drives
(photo Bosch Rexroth).
The Stuttgart SmartShell on the Vaihingen campus of the
University of Stuttgart (photo: Bosch Rexroth) has hydraulic
cylinders that compensate for extreme loads on the structure.
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
SEITE
29 PAGE
29
Control concept for ultra-lightweight structures
Brains
Architecture
with
Ultralight, Adaptive Structures
Trafc, snow and wind are typical,
everyday loads that buildings are
subjected to. In contrast, extreme
events such as earthquakes or the
worst winter in living memory are
very rare yet buildings still have
to be designed to cope with them.
Today, this means using large quan-
tities of materials to ensure safety
margins that are hardly ever needed.
So that future buildings can use less
material but still cope with extreme
events, the University of Stuttgart
is investigating ultra-lightweight,
adaptive structures with intelligent
hydraulics that compensate for a
wide range of different loads.
Static Adaptation and Active
Vibration Control
The Stuttgart researchers are study-
ing both static and dynamic forces
and their effects. Static forces are
due to factors such as snow, while
dynamic forces (i.e., vibrations) are
caused by factors such as wind
gusts. To reduce the resulting stresses,
the load situation rst has to be
captured by sensors. The control
algorithm then uses the measure-
ment values to compute the opti-
mum control for the actuators that
adjust the structure to the external
loads. For static forces, the struc-
ture is adapted statically. If vibra-
tions need damping, however, the
sensors have to continuously cap-
ture the stress and control the actu-
ators dynamically. As a result, load-
bearing elements can be smaller
and lighter: The structural mass
saved in the construction is replaced
by short-term energy inputs.
The Stuttgart SmartShell
Prototype
To test their control concepts, the
Stuttgart university institutes worked
together with Bosch Rexroth to con-
struct the Stuttgart SmartShell
prototype, the rst adaptive load-
bearing shell structure in the world.
Constructed from wood in a four-
layer laminate, it is only 40mm
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
SEITE 30 PAGE 30 UNIVERSITY OF STUTTGART
model and on feedback control that
is computed from the model and
used in the Stuttgart SmartShell.
Because the feedback control is
derived directly from the analytical
model description, it can be adapted
to the time-variable parameters of
the shell, such as the static load sit-
uation or the set points for the sup-
port point positioning. The feed-
back control is designed to optimize
the damping properties of the con-
trol while at the same time minimiz-
ing the energy required for damp-
ing. Both these adaptivity require-
ments can be met simultaneously
(gures 4, 6) by combining the set
points for the static adaptation and
active vibration control (gure 3).
Prototype Implementation
The control system (gure 5) of
the Stuttgart SmartShell essentially
has to:
Control the hydraulics
Cyclically monitor the status and
safety
Read the measurement data from
the cylinders and strain gauges
Estimate the vibration state and
reconstruct the static load
Generate the set points for the
static adaptation and vibration
control
Perform high-precision actuator
control
The system that performs these com-
plex tasks consists of a dSPACE
DS1006 Processor Board and a
Motion Logic Control (MLC) unit
from Bosch Rexroth. While the MLC
performs the control tasks for the
hydraulic system and sends the con-
trol inputs for positions and forces
to the individual hydraulic drives,
the DS1006 estimates the state and
generates the static and dynamic
set points. The DS1006 also com-
putes the direct and inverse kine-
matics. These are used to convert
the set points for the support point
thick. Comparable struc tures built
by traditional, passive methods
require several times as much mate-
rial. The shell weighs 1.4 metric tons
and rests on four supports, three
of which can be moved by hydrau-
lic cylinders.
Each of the active supports is
mounted on a tripod of actuators
(gure 2) that allow it to be posi-
tioned freely in three-dimensional
space. The kinematic relationships
between the lengths of the hydrau-
lic cylinders in the tripod arrangement
and the positions of the supports
are continuously computed online.
A wide range of sensors are available
for measuring the system state. For
example, the hydraulic cylinders are
Figure 4: Comparison between undamped vibration (blue curve) and actively controlled
vibration (red curve). The damping control is activated at 3s and reduces the duration of
vibration by 80%.
Figure 3: Control concept for static adaptation and active vibration control.
equipped with high-resolution dis-
placement sensors and force sen-
sors. Strain gauges mounted on the
shell structure capture the strain at
specic mea surement points.
The Control Concept
In the rst step, simulation models
are used to determine the necessary
static adaptation. The models repre-
sent the behavior of the structure
under different loads, and the opti-
mum positions of the supports are
calculated from that. To dampen
vibrations with maximum efciency,
a description of the dynamic behav-
ior was added to the model. The
control concept for vibration damp-
ing is based on a dynamic behavior
Controller
Set point
generator
Adaptation
Modal
observer
Reconstruction
of
disturbance
d
q* e u
AVC u
y
d u
STA
Static Adaptation
Active Vibration Control
q
Shell
structure
0 5 10 15
-100
0
100
Time (s)
S
t
r
a
i
n
(
m
/
m
)
AVC inactive
AVC active
AVC = Active Vibration Control
AVC off AVC on
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
SEITE
31 PAGE
31
Figure 5: Software and hardware structure for implementing the static and dynamic adaptation.
Supporting structure for assembling the Stuttgart SmartShell.
With the dSPACE prototyping-system, adaptivity concepts for ultra-light-
weight structures can be implemented in a short time.
Martin Weickgenannt, University of Stuttgart
Adaptation and Vibration Control
DS1006MC
Observer/controller
Sample time: 1 ms
Set point lter
Sample time: 5 ms
Read strain gauge
Sample time: 2 ms
Kinematics/inverse kinematics
Sample time: 5 ms
dSPACE ControlDesk
Strain
gauge1
Strain
gauge2
Strain
gauge14
Drive Control
MLC L65
Process control
Cycle time: 10 ms
Process monitoring
Safety functions
Drive control
Cycle time: 8 ms
Implement position
set points
Operating console VEP50
Probus
C
A
N
Sercos III
Support3
Cylinder, valves
Support2
Cylinder, valves
Support1
Cylinder, valves
Power unit
Pressure supply
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
SEITE 32 PAGE 32 UNIVERSITY OF STUTTGART
Figure 6: Stress distribution due to an exemplary external load. Adaptation reduces peak stress
by 66%.
Conclusions
Using a dSPACE controller board
together with a Rexroth Motion
Logic Control (MLC) combines
the advantages of a rapid proto-
typing system and a process
ECU. This means that the tasks
of the Stuttgart SmartShell
generating static and dynamic
set points, plus process control
can be implemented efciently.
positions and forces into local set
points for the drives. The current
values for the positions and forces
at the support points are obtained
from the local measurements, which
are then transformed into the global
coordinate system. The high com-
plexity of the kinematic calculations
and the short cycle time of 5ms
involve a large computation load.
The convenient conguration
options of the DS1006 are particu-
larly helpful here, as the computa-
tion processes can easily be distrib-
uted across its processor cores to
ensure optimum use of the process-
ing capacity. The Probus connec-
tion between the DS1006 and the
MLC allows the high-frequency
exchange of the set points, mea-
surement values and control com-
mands at a rate of 12MBit/s. The
strain sensors are connected to the
DS1006 via three separately con-
trolled CAN networks, providing a
sampling rate of 200 Hz. The CAN
interface can be set up quickly and
conveniently by means of the RTI
CAN MultiMessage Blockset.
Finally, with the experiment software
Snow, wind and trafc are typical forces that structures have to withstand.
y
x
y
x
y
x
y
x
Passive state,
top surface of shell
Passive state,
bottom surface of shell
Adaptive state,
top surface of shell
Adaptive state,
bottom surface of shell
57 MPa 0
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
SEITE
33 PAGE
33
Oliver Sawodny
Oliver Sawodny is the head of the Insti-
tute for System Dynamics at the Uni-
versity of Stuttgart.
Stefan Neuhuser
Stefan Neuhuser is a research associate
at the Institute for Lightweight Structures
and Conceptual Design at the University
of Stuttgart.
Martin Weickgenannt
Martin Weickgenannt was a research
associate at the Institute for System
Dynamics of the University of Stuttgart
and is now with Drr Systems GmbH.
ControlDesk, it is easy to specify
items such as set points for the
individual drives and to switch
between different control strate-
gies.
Martin Weickgenannt
Stefan Neuhuser
Werner Sobek
Oliver Sawodny
University of Stuttgart
The shell structure is only 40 mm thick.
Werner Sobek
Werner Sobek is the head of the Insti-
tute for Lightweight Structures and
Conceptual Design at the University of
Stuttgart and founder of the Werner
Sobek group.
Watch this video to see the
Stuttgart SmartShell in action:
www.youtube.com/watch?v=vDb2h1-7LA0
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
UNIVERSITY OF KAISERSLAUTERN PAGE 34
Limited resources, climate change and a rapidly growing world population
make efciency and sustainability the biggest challenges in horticulture and
agriculture. Using the latest technologies is the key to meeting these chal-
lenges. Year for year, a team of students from the Institute of Mechatronics
in Mechanical and Automotive Engineering at the University of Kaisers lautern,
and many other teams, participate in the Field Robot Event, an international
competition for autonomous eld robots, giving a detailed picture of the
new generation of agricultural technology.
Robots
in the
Rose Garden
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE
35
in the
Rose Garden
Autonomous Vehicles for Horticulture
and Agriculture
Autonomous Vehicles for Horticulture
and Agriculture
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
A Small Revolution
Intelligent sensors, GPS-based auto-
matic steering systems with centimeter-
level accuracy, robots and autono-
mous vehicles are the next major step
in agricultural engineering. To meet
these enormous technological chal-
lenges, new creative approaches and
interdisciplinary thinking are needed.
The Institute of Mechatronics in
Mechanical and Automotive Engi-
neering (MEC) at the University of
Kaiserslautern set up a team of stu-
dents to take on this task. Together,
the students developed an autono-
mous vehicle for eld use.
Figure 1: A task of the Field Robot Event 2012 was to drive autonomously through rows of plants in order to complete a given route.
UNIVERSITY OF KAISERSLAUTERN PAGE 36
Because dSPACEs MicroAutoBox II plus Embedded PC combines the
proven tools of hardware-related ECU development with the exibility
of a PC, it was an excellent platform for our autonomous vehicle.
Roland Werner, University of Kaiserslautern
The Competition
The highlight of the teams year on
the project was participating in the
Field Robot Event, an annual inter-
national eld robotics competition,
held in 2012 in Venlo (Netherlands)
as part of the Floriade (international
garden show). The overall task of
the Field Robot Event was to let a
vehicle navigate autonomously
through rows of plants (gure 1).
Missing plants, and curved and
blocked rows made it difcult for
the vehicles to move along. On
their way, the vehicles were re-
quired to perform additional tasks
related to agriculture. The assess-
ment of the teams was based on
the distance traveled in 3 minutes,
with bonus points for completed
tasks and penalty points for plant
damage and manual intervention.
A particular challenge of last years
Field Robot Event was nding, pick-
ing up and returning a randomly
chosen, specially marked plant among
rows of rose pots.
The World from the Vehicles Eye
A large number of sensors give
vehicles the ability to recognize their
environment (gures 2 and 3). One
S - 3R - 2R - 1L - 0 - 1L - F
Start
Finish
max. 1 m
max. 1 m
max. 1,5 m
75 cm
30-45 cm
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE
37
important component for autono-
mous driving is a set of two laser
scanners mounted on the front and
back of the vehicle. Delivering a
240 eld of view in one scanning
plane, the scanners allow to detect
obstacles in the vehicles surrounding
environment. For the turning at the
rows end, the yaw rate measured
by inertial sensors and the resulting
calculated vehicle orientation is cru-
cial. Located on the vehicle roof, six
Figure 2: The main components of the autonomous vehicle developed by students from the Institute of Mechatronics in Mechanical and
Automotive Engineering (MEC) at the University of Kaiserslautern.
stereo webcams with a total 360
eld of view allow to nd the
marked plant and provide informa-
tion on the vehicles distance to the
plant. In addition to these sensors for
autonomous navigation, the vehicle
carries a WLAN interface and an R/C
receiver for development and testing
purposes.
Four-Wheel Steering and Drive
Unpaved surfaces, obstacles and
narrow turns between rows are
challenges for the vehicles actua-
tors. Four servo motors allow for
steering all wheels individually.
A central brushless DC drive with
2 kW power and a permanent four-
wheel drive provide the necessary
force at the four tires. An intelligent
power converter supplies the drive
from the vehicles 33 V LiFePO bat-
tery pack.
DC/DC converter 5V and 12V
Brushless DC drive controller
33V LiFePO battery packs
WLAN MicroAutoBox II with Embedded PC
Inertial sensor
Laser scanner
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
UNIVERSITY OF KAISERSLAUTERN PAGE 38
Figure 3: Block diagram of the autonomous vehicles sensors, signal processing and actuators.
Figure 4: Object recognition via OpenCV with a marked plant in the original image (left: cylinder in red, blue and yellow), recognized color
regions (right: rectangles in red, blue and yellow) and identied plant marking (right: white rectangle).
Distributed Signal Processing I:
Real-Time Control with
MicroAutoBox II
For signal processing and actuator
control, the vehicle uses the advan-
tages of two worlds. Hardware-
parts of the software running on
MicroAutoBox II. The multitude of
I/O interfaces makes it easy to con-
nect to all different kinds of sensors
and actuators (gure 3). MATLAB
/
Simulink
and Linux.
The combination of MicroAutoBox II
and Embedded PC make integrating
the image processing tasks into the
overall system quite easy. The inter-
nal Gigabit Ethernet switch and
UDP/IP guarantee rapid communica-
tion between MicroAutoBox II and
Embedded PC running Linux. One
particular challenge of the Field
Robot Event 2012 was to have the
autonomous robot identify a ran-
domly chosen, specially marked rose
pot and stop in the immediate vicin-
ity of the pot. The marking selected
for the pot was a cylinder with
Conclusion and Outlook
The Field Robot Event 2012 was the
second stage for the development of
the MEC vehicle. In comparison to
its debut the year before, the team
improved its placement, moving from
10th to 6th place out of 18 teams.
In particular, the platform shift to
dSPACE MicroAutoBox II was a posi-
tive improvement. The reliable, pro-
ven development tools allowed the
team to focus their attention on
function development. Furthermore,
the combination of MicroAutoBox II
and Embedded PC was especially
benecial because the team could
streaks of red, blue and yellow
(gure 4). Recognizing the mark-
ings is performed sequentially, with
the rst step being the identica-
tion of image areas containing a
high proportion of the markings
colors. Next, the arrangement of
these areas is used to differentiate
between the searched markings
from other random image compo-
nents in the environment. The size
of the discovered markings and the
depth information of the stereo im-
age serve as the base values for
measuring the distance between
the vehicle and marked plant.
Roland Werner
University of Kaiserslautern
implement the hardware-oriented
applications with real-time require-
ments and a variety of I/O interfaces
without having to sacrice the exi-
bility of a PC. The next step for
the students from the University of
Kaiserslautern will be to enter their
further optimized vehicle at the Field
Robot Event 2013, taking place at
the Czech University of Life and
Environmental Sciences (CZU) in
Prague.
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Perfect
Balance
Controlling Li-ion cell voltages
during prototyping
To benet fully from the
high energy density of
Li-ion batteries, the state
of charge of the individual
cells must be monitored
precisely. dSPACE has
developed a battery man-
agement system that per-
forms this task throughout
the development process,
from the rst model to
in-vehicle testing. Its main
focus is on measuring and
controlling Li-ion batteries.
Electromobility Needs Powerful
Batteries
One reason for the combustion
engines great success in the 20th
century is gasolines high energy
density. While one liter of gasoline
can run for many kilometers, a
modern battery of the same mass or
volume takes an electric vehicle only
a fraction of the distance. As this
comparison clearly shows, develop-
ing powerful, high-density batteries
is key to the breakthrough of elec-
trical vehicles.
Li-Ion Batteries as the Solution
One type of high-density power
storage is batteries based on Li-ion
technology. A Li-ion battery consists
of numerous single cells with voltages
in a typical range of 3.3 4.2 V,
depending on the battery type. To
reach the necessary voltages of sev-
BATTERY MANAGE MENT SYSTEM PAGE 40
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Balance
PAGE
41
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
With the battery management system from
dSPACE, high-voltage batteries can be man-
aged at cell level with high accuracy.
health (SoH), and so on. The BMS
also performs what is known as cell
balancing to keep the voltages of
all the single cells at the same level.
In passive balancing, superuous
charge is drawn off by switching an
ohmic resistance in parallel to any
cells whose voltage is too high. The
voltage balance between the cells in
a stack is one of the most important
factors that inuence the life span
of Li-ion batteries.
The Battery Management System
from dSPACE
Because standard solutions fail to
meet the requirements described
above, at dSPACE we developed our
own battery management system
for rapid control prototyping, called
Battery Cell Voltage Measurement
and Balancing. This system lets users
measure and control the cell volt-
ages of Li-ion batteries during rapid
control prototyping, and also sup-
ports the development of battery
management algorithms.
Li-ion batteries with a total voltage
of up to 846 V can be connected to
this dSPACE BMS (battery manage-
ment and balancing system). The
system measures the voltage of
each cell with high precision and
provides the ability to reduce it, if
Figure 1: Typical load curve of a Li-ion cell. The usable voltage range is limited to several 100 mV.
Figure 2: The modular structure of the dSPACE battery management system allows tailor-
made congurations of up to around 200 cells and can also be installed directly in a vehicle.
eral hundred volts, a large number
of single cells are combined to build
a cell stack.
Li-Ion Batteries Need Monitoring
Li-ion batteries have to be constantly
monitored and controlled because
the usable voltage range of a Li-ion
cell is limited to several 100 mV (g-
ure 1). The further the voltage moves
out of this ideal range, the more the
life span of the cell is impaired. In
extreme cases, the cell can even be
destroyed. Instances of battery res
in telephones, laptops, and last but
by no means least, the Boeing Dream-
liner, emphasize just how important
it is to monitor the battery state.
Protecting the cells against under-
voltage and overvoltage is therefore
a top priority. This is where a battery
management system (BMS) comes
in. This system has the challenging
task of measuring cell voltages with
galvanic isolation at a precision of
only a few millivolts, when the volt-
age of the overall cell stack usually
reaches several hundred volts. A
BMS also has to monitor the tem-
perature, because this has a major
impact on the cell state.
Using algorithms, the BMS continu-
ously monitors the individual cells
to determine values such as the bat-
terys state of charge (SoC), state of
Usable range
4,5
4,0
3,5
3,0
2,5
2,0
0
20 40
State of charge (%)
V
o
l
t
a
g
e
(
V
)
60 80
Under-
voltage
Over-
voltage
BATTERY MANAGE MENT SYSTEM PAGE 42
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
dSPACE
PGI1
Ethernet
Ground
dSPACE
system platform
R
F
Isolation
watchdog
Sensor
Board
Board with
balancing
resistances
Cut-off
device
Battery
Electrical
isolation
Enclosure Isolation fault
Features at a Glance:
Max. no. of cells Approx. 200
Max. voltage 846 V
No. of cells per module 24
Cell voltage 0 5 V
Measurement frequency Max. 1 kSPS
Precision 3 mV at 3.3 V 300 mV
Resolution 0.61 mV
Synchronized measurement For all cells
BMS IC Intersil ISL78600
Temperature measurement For each cell, by thermistor (NTC)
Balancing method Passive balancing
Isolation Complete galvanic isolation
Isolation monitoring Interface for isolation monitoring device
Safety features Watchdog, diagnostics options, temperature monitoring
Hardware interface Ethernet
Software interface Simulink
blockset
ages and switch-off times so that
developers can give their full atten-
tion to the more important BMS
algorithms.
Seamless System Monitoring
To continuously monitor the overall
system state, the dSPACE BMS has
extensive error detection and alarm
features:
Warnings on hardware faults and
communication/synchronization
errors
Temperature warnings
Warnings on isolation faults
Warnings on cell undervoltage
and overvoltage
To ensure the safety of the overall
system, the isolation resistance is
monitored by a separate device that
constantly provides information on
the state of the isolation and imme-
diately outputs an alarm when it
detects an isolation leak.
For safety reasons, dSPACE supplies
the BMS only as a turnkey system
in an engineering project.
required, by means of a switchable
resistor.
The system is modular and can be
assembled to create congurations
of between 6 and approx. 200 cells.
It can also be installed directly in a
vehicle. The modules are connected
via Ethernet with a dSPACE proto-
typing system such as MicroAuto-
Box II.
One module can measure the voltage
and temperature of up to 24 cells
and passively balance each cell sep-
arately. The balancing resistances
are located on a separate carrier
board so that different resistance
values can be tried out quickly.
Because the voltage and tempera-
ture measurement inputs are gal-
vanically isolated, a real battery can
also be connected to the system.
All the properties of a prototyping
system have to be more powerful
than those of the nal product. The
dSPACE BMS achieves a precision of
3 mV at a measurement frequency
of up to 1 kHz, regardless of the
number of cells. This makes it pos-
sible to study the cell chemistry as
well. The cell voltage measurement
processes in the individual modules
can also be synchronized.
In manual balancing mode, the
system gives users all the freedom
they need to develop battery man-
agement algorithms, letting them
balance cells individually or collec-
tively, and at any intervals. The `auto-
matic balancing mode is a comfort
function that performs the less im-
portant ask of specifying target volt-
Figure 3: The isolation concept of the dSPACE battery management system makes it safe
to use high battery voltages.
PAGE
43
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
CONTROLDESK NEXT GENERATION PAGE 44
As of version 4.3, ControlDesk Next Generation (dSPACEs experiment soft-
ware) provides major innovations for automotive, avionic and industrial
applications. For example, camera images can be recorded with a video
instrument via an integrated camera interface and replayed in synchronicity
with measurement recordings. And for avionics, there is a new Primary
Flight Display (PFD), with Altimeter, Articial Horizon, Heading Indicator
and Airspeed Indicator instruments.
The video instrument and camera interface are used to record and replay video and bus data synchronously (shown here is ControlDesk 5.0
with its new user interface).
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE
45
A
Clear View
New ControlDesk instruments for automotives,
aviation and industry
Video Instrument and Camera
Interface
There are numerous applications
where it is useful to record camera
images synchronously to the mea-
surement values displayed in Control-
Desk
/
Simulink
and support all the essen-
tial simulation tasks in the electrics/
electronics development process.
dSPACE provides models for combus-
tion engines, electric components,
vehicle dynamics and driver assis-
tance systems. We rst launched
them on the market in 2005, and
have been extending them continu-
ously ever since.
What were the goals, and what
has been achieved so far?
From the beginning, we have had
two main goals: One was to give
our customers the option to obtain
all the components needed for HIL
simulation from a single vendor, since
this provides enormous conguration,
operation and maintenance benets.
And the other was to focus on mak-
ing the simulation process more
Winter Tests
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
ASM Model
Portfolio
Models for combustion
engines:
ASM Diesel/Gasoline Engine
(mean-value models)
ASM Diesel/Gasoline InCylinder
(physical models)
ASM Turbocharger (physical
model)
ASM Diesel Exhaust System
(physical model)
Vehicle dynamics models:
ASM Vehicle Dynamics
ASM Truck
ASM Trailer
ASM Brake Hydraulics
ASM Pneumatics
ASM KnC
Environment model:
ASM Trafc
ASM Environment
ASM INTERVIEW PAGE 48
What special features do the models
have?
One strong point is that they have
standardized interfaces that let users
combine models like building blocks,
to tailor them precisely to the scope
and quality required for each spe-
cic application. For example, they
can put together models that repre-
sent a hybrid drive with a specic
battery cell characteristic.
ments at Daimler in Stuttgart, Nissan
in Yokohama, and Scania in Sweden.
Our models are also used by GM in
America, Volkswagen in Europe, and
Mazda in Asia, to name but a few. The
heavy vehicle sector is represented by
Caterpillar, Volvo, WABCO and oth-
ers. These companies particularly ap-
preciate our powerful variant man-
agement, which helps them param-
eterize and simulate vehicle variants.
This is easy to do in the graphical
tool ModelDesk, which is another
boost to productivity. Another
strong point is that the models
match the dSPACE hardware very
well. This supports the emulation
of cell voltages, the simulation of
electrical drives for process-based or
FPGA-based calculations, etc. etc.
Which customers use the models?
Even though the market launch of
the models was only a few years
ago, there are now 800 licenses at
250 customers all over the world.
This especially includes major auto-
mobile manufacturers and suppliers,
and also several universities. Recent
newcomers are important depart-
Why do so many companies rely
on ASM?
One reason is that the ASMs provide
a complete portfolio from engines
to trafc models, enabling entire
virtual vehicles to be simulated
together with their environment.
dSPACE is currently the only vendor
to offer such a comprehensive range
of models. Another aspect is that
many of our customers no longer
want to do all the maintenance and
development work involved in using
their own in-house models. Chang-
ing to the ASMs saves them a lot of
work. At the same time they remain
exible enough for customer-specic
adaptations, thanks to the openness
of the ASM. And whenever their
dSPACE plans to strategically extend the range
of models provided for electromobility and
driver assistance systems.
ASM Video Channel
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE
49
developers face new simulations
tasks, they can either quickly nd
whatever they want from our well-
established model portfolio, or fur-
ther develop the models with our
support.
What application elds will the
ASMs be tackling next?
The main ones will be electromobility
and driver assistance systems. Our
customers are very active in these
elds, and we will support them
with simulation models. The ASM
Trafc Model is already facilitating
the development of ACC, emer-
gency braking, lane keeping and
parking assistants. Car2x is also a
very exciting development eld, and
we will be launching simulation
tools for this some time this year.
Simulating trafc at junctions, map-
based systems, and object recogni-
tion are just a few examples. The
key to all of this is a exible, conve-
nient environment simulation this
is where ModelDesk comes in.
In the eld of electric drives, we will
be adding more precise component
models to our portfolio, including
FPGA-based ones. We are also ad-
vancing classic application elds like
combustion engines and vehicle
dynamics simulation, for example,
by further developing high-resolution
cylinder pressure simulation for
combustion-based engine controls
and enhancing ASM KnC as a virtual
suspension test bench.
How are you preparing for the
development challenges of the
future?
Electrics/electronics systems are
growing in functionality and also
becoming more highly networked.
As a result, larger and larger models
have to be combined. This leads to
new challenges regarding not only
the contents but also the handling
of the models. To facilitate handling,
we will provide highly integrated
data management. Connection to
dSPACEs data management and
collaboration software SYNECT
will offer powerful solutions dedi-
cated to ofine simulation.
How do you acquire the know-how
for simulations?
When developing the ASMs, we
cooperate closely with our users,
so whenever new requirements
come up, we can include them in
our development plan immediately.
Some parts of models that were
developed in customer projects go
straight into the product. And for
specic tasks, we work with part-
ners who already have the relevant
experience. In the important appli-
cation elds of electromobility and
driver assistance systems, we are
actively involved in the Toolbox
Speichersysteme, TRAFFIS and Vir-
tual Car2x research projects. Well-
founded research results coupled with
dSPACEs practical experience and
modeling expertise are the ideal
foundation for powerful simulation
models to perform the simulation
tasks of the future.
Thank you for speaking with us,
Dr. Haupt.
Dr. Hagen Haupt is in charge of simulation model deve-
lopment at dSPACE GmbH.
ECU
Simulink/VEOS
Automotive Simulation Models
Frontloading tests
EGO vehicle
Trafc
Sensors
Driver
Street
The ASMs enable early, integrated testing in the automotive E/E development process.
dSPACE is the only model vendor who covers
the entire scope of automotive simulation from
engines to road trafc.
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
With the new Ethernet Conguration Package, dSPACE is the rst provider
to enable service-based communication in Ethernet ECU networks in real-
time systems.
New Ethernet Conguration Package makes
network conguration easier.
ETHERNET CONFIGURATION PACKAGE PAGE 50
net
-
working
Easy
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Why Automotive Ethernet?
Following FlexRay's successful intro-
duction into vehicles just a few years
ago, Ethernet is now about to go
into production use as an automotive
communication bus. With its exible
layer model, high bandwidth and
cost-effective, independent imple-
mentations, Ethernet has numerous
potential uses in vehicles. Ethernet
networks will therefore play a deci-
sive role in modern driver assistance
PAGE 51
dSPACE and Ethernet
For many years, dSPACE has been
providing products and solutions for
connecting Ethernet to real-time sys-
tems. Two examples are the DS1006
Processor Board for hardware-in-the-
loop simulation and the MicroAuto-
Box for rapid control prototyping with
blocksets designed for MATLAB
/
Simulink
.
These products focus mainly on the
User Datagram Protocol (UDP) and
Ethernet networks will therefore play a decisive
role in modern driver assistance systems.
systems, new comfort and entertain-
ment functions, ECU ashing, and
other functions. Because the Ether-
net networks have often been opti-
mized for automotive use, develop-
ment is also focusing on many other
aspects of the communication system.
For example, the OPEN Alliance Spe-
cial Interest Group (OPEN Alliance
SIG), which dSPACE is also a member
of, is promoting the widespread intro-
duction of Ethernet-based networks
with single pair, unshielded twisted
cables. Activities in the elds of
ASAM MCD-2-NET (FIBEX) and AUTO-
SAR include ones aiming to standard-
ize communication descriptions and
harmonize middleware layers. There
is already a broad base of users and
interested parties who are planning
to introduce automotive Ethernet.
dSPACE supports these customers
with the new Ethernet Conguration
Package.
the Transmission Control Protocol /
Internet Protocol (TCP/IP) via Ether-
net. However, the current Ethernet/IP
discussion is addressing the layers
above UDP and TCP/IP and aims to
implement service-based communi-
cation (gure 1). The serialization
protocol SOME/IP for service-based
communication and the service dis-
covery protocol SOME/IP-SD are
playing a key role here.
A version of the FIBEX standard
called FIBEX4, published by ASAM
[www.asam.net] back in September
2011, meets the need to describe
the additional elements required for
service-based communication via
Ethernet. The experience gathered
from the initial application phase will
lead to FIBEX 4.1, which is due to be
released soon. FIBEX 4 is the rst
step towards a data exchange for-
mat for Ethernet-based in-vehicle
communication networks. More
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
the bex4it schema and elements
for the IP and transport addresses
(such as ports) were added. The FIBEX
specication was extended by two
basic concepts for communication
via Ethernet. The rst of these is
typical signal-based communication
like that used with CAN. This involves
mapping protocol data units (PDUs)
to Ethernet. The network and appli-
cation endpoints for the PDUs are
then modeled in FIBEX. Because
Ethernet provides more features
than just transmitting simple signals
via UDP, etc., the second basic con-
cept allows the specication of com-
plex service interfaces that contain
methods for describing communica-
tion on higher layers. Generic data
types can be used for the param-
eters of the methods to transmit
information that is more structured
than with simple signals. The bex-
4services schema contains the ele-
ments for modeling service-based
communication. Underneath the
application endpoints are the ser-
vice interfaces that are instantiated
as either provided services or con-
sumed services.
Tool Support
The dSPACE Ethernet Conguration
Package supports the simulation of
Figure 1: Example of service-based ECU communication
ETHERNET CONFIGURATION PACKAGE PAGE 52
Ethernet Congura-
tion Package
Features
Enables the simulation of
service-oriented, event-based
Ethernet communication in
real-time systems.
Supports FIBEX4 communica-
tion descriptions.
Supports SOME/IP middle-
ware.
communication descriptions for
service-based communication via
Ethernet will follow, especially ones
for AUTOSAR. The main idea behind
this IP- and service-based commu-
nication is explained below with
excerpts from the FIBEX4 format.
Service-Based Communication
with FIBEX4
Ethernet is specied in IEEE 802.3
and covers the rst two layers of
the ISO/OSI model. Because the
network nodes share the physical
transmission medium, collisions
can occur. These can be avoided
by point-to-point connections with
switches as coupling elements. The
bex4ethernet schema extends the
FIBEX topology with the elements
that are needed for describing Ether-
net in the data link and physical lay-
ers. Network endpoints and applica-
tion endpoints were integrated into
Restbus simulation
ECU Central-
Control Unit
ECU Light Unit
Provided Service Instance
GPS
Network
Endpoint
172.29.177.1
Application Endpoint 30491
Application Endpoint 30477
Application Endpoint 30489
Consumed Service Instance
POI
Consumed Service Instance
Light
Application Endpoint 30477
Application Endpoint 30489
Application Endpoint 30492
Network
Endpoint
172.29.177.3
Network
Endpoint
172.29.177.2
ECU Display-
Control Unit
Provided Service Instance
POI
Consumed Service Instance
GPS
Provided Service Instance
Light
Consumed Service Instance
GPS
dSPACE Simulator Time Sync
ISO/OSI layer model of service-based communication via Ethernet
Control Communication Service Discovery
IEEE
802.1AS
UDP and/or TCP
IEEE Ethernet MAC + VLAN (IEEE 802.1Q)
Automotive Ethernet Physical Layer (BroadR-Reach)
IPv4
SOME/IP
SOME/IP-SD
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
service- and event-based communi-
cation as described in FIBEX4. The
rst version of the dSPACE Ethernet
Conguration Packages is available
for systems with quad-core DS1006
Processor Boards. Like the dSPACE
FlexRay support, the dSPACE Ether-
net Conguration Package consists
of two parts: One is the dSPACE Eth-
ernet Conguration Tool for cong-
uring a dSPACE system as a simulation
node in a service-based Ethernet
network. The other is the RTI Ether-
net Conguration Blockset for mod-
eling service-based communication
in MATLAB/Simulink (gure 3).
The Ethernet Conguration Tool
A FIBEX4 le can be imported and
visualized with the dSPACE Ethernet
Conguration Tool. FIBEX elements
such as clusters, ECUs, services and
events are represented in a clearly
organized tree (gure 2). Users can
select ECUs, services and events for
the simulation simply by drag &
drop. A structured view displays the
most important attributes of the FIBEX
elements. The selected services and
events serve as inputs to automatic
communication code generation. A
transmission le is generated as a
basis for service-based model frame
generation in MATLAB/Simulink. The
Figure 3: Architecture of the Ethernet Conguration Package
Figure 2: Ethernet Conguration Tool
PAGE 53
Simulation on dSPACE hardware
dSPACE Ethernet
Conguration Tool
RTI Ethernet
Conguration Blockset
Conguration of services and
events for simulation
Ethernet Blockset for modeling
with MATLAB
/Simulink
FIBEX
4.0
*.m
*.c
result is a Simulink interface model
with precongured service and event
blocks that build on the RTI Ether-
net Conguration Blockset. The
Simulink interface model provides
subsystems for the simulated ECUs
with the relevant service instance
interfaces. With the help of the
parameterized event blocks, users
can design functions in the func-
tion model, for example, to try out
the new communication via Ether-
net/IP or to perform restbus simula-
tion. The new dSPACE Ethernet
Conguration Package is the rst tool
for restbus simulation on real-time
systems for service-based in-vehicle
Ethernet communication.
The latest version, Ethernet Congu-
ration Package 1.1, was released in
June and also supports dynamic
service discovery. Further versions
will follow.
Start code generation Selected nodes, services
and events for restbus
simulation
View of important
element attributes
Log view
Visualization of FIBEX
elements, i.e., clus-
ters, nodes, services
and events
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
Success
SCALEXIO in
customer projects
SCALEXIO CUSTOMER PROJECTS PAGE 54
Launched two years ago, the SCALEXIO hardware-in-the-loop (HIL) test sys-
tems from dSPACE have already proven their value in numerous customer
projects. They are now being used across the globe, in countries such as
China, France, Germany, Italy, Japan, the UK and the USA.
Broad Range of Applications
SCALEXIO
(ABB,
Audi) and VEOS
(Volkswagen). And
as several presentations demonstrated,
numerous customers have already
given SCALEXIO
or
SIL/MIL/PIL testing in the context of
Simulink
and TargetLink
). Clear,
well-organized analyses of tests, test
coverage and test progress can be
generated quickly and simply by
using the test results stored centrally
in SYNECT, and then used for plan-
ning further tests. The module can
be combined with SYNECT Variant
Management for close interaction
between variant management and
test management.
The latest module is SYNECT Signal
& Parameter Management. This
gives users the capability to manage
signals, parameters and parameter
sets centrally throughout the entire
development process. SYNECT Signal
& Parameter Management supports
numerous le formats for exchang-
ing signals and parameters and
using them with other tools in the
development process. For example,
Simulink users can exibly exchange
parameter values, and signal and
parameter denitions, between their
function or plant model and SYNECTs
central data management facility.
TargetLink users can exchange vari-
ables, scalings and type denitions
between the le-based TargetLink
Data Dictionary and SYNECT. Param-
eters for ECU calibration can be im-
ported and exported in DCM and
PAR les. With SYNECT Variant Man-
agement, users can specify depend-
cies between parameters and variants,
and create and manage parameter
sets for different variant congura-
tions.
Yet another SYNECT module, for
model management, is planned for
the end of 2013.
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
PAGE B PAGE
61
MicroAutoBox II with
Multistage Watchdog
Mechanism
dSPACE has now equipped its
MicroAutoBox
II with a multistage
watchdog mechanism. Its run-time
monitoring, with congurable time-
out behavior, yields a higher level of
security for prototype vehicles used
in road and eet tests because the
system can immediately respond to
critical situations appropriately and
as dened by the customer. The sys-
tem watchdog is implemented via
FPGA in the hardware independently
of the MicroAutoBox II processor,
while the individually congurable
software watchdogs enable targeted
monitoring of the software tasks in
the user model. As soon as the user
software generates a timeout in the
system watchdog, the congured
behaviors kick in. This timeout
behavior includes examples such
as backing up data, setting dened
output values, or a complete system
reboot. The new watchdog mecha-
nism comes with the dSPACE 2013-A
software release. For simple, block-
diagram-based programming and
integrating the new watchdog into
existing Simulink control models,
dSPACE provides a new dSPACE
Real-Time Interface Blockset (RTI)
as an interface. This new Watchdog
Blockset supports all MicroAutoBox
II variants. At least 20 tasks (more is
possible, depen ding on the memory
resources being used) can be moni-
tored simul taneously and indepen-
dently from each other. In addition,
individual timeout behavior can be
assigned to each respective task.
Implement FPGA Models Faster with
XSG Utils Library
The XSG Utils Library from dSPACE
gives users of real-time FPGAs a
large number of extended function
blocks for implementing their own
projects. These range from extended
I/O, scope functions and look-up
table functions to an average calcu-
lator, a sine generator and a wave-
table encoder. The function blocks
can be used in rapid control proto-
typing projects with dSPACE Micro-
AutoBox
II
1401/1511/1512 with an inserted
DS1552 Multi-I/O Module can now
be used for FPGA-based cylinder
pressure measurement and indica-
tion (MABXII Cylinder Pressure Indi-
cation Solution). The solution sup-
ports angle-synchronous pressure
measurement with 0.1 max. resolu-
tion. The evaluation of the pressure
values can be controlled by several
algorithms in real time, such as
released heat, mass fraction burned,
the indicated mean effective pres-
sure, etc.
SCALEXIO Solutions for
Aerospace Applications
dSPACE has added the rst I/O inter-
faces for aerospace applications to
its SCALEXIO
hardware-in-the-loop
(HIL) system. The new interfaces
provide access to ARINC-429 and
MIL-STD-1553 networks, and are
based on tried-and-trusted PMC/
XMC modules from Avionics Inter-
face Technologies (AIT). These mod-
ules use the PCI(e) interfaces in the
SCALEXIO Processing Unit to en-
sure the optimum bandwidth for
data exchange with the real-time
model. To guarantee real-time capa-
bility, dSPACE developed dedicated
QNX drivers for this application
case.
Please let us know your opinion on the quality of the
dSPACE Magazine. Just ll out the enclosed reply card
and return it to us. You can also use the card to order
any other information by post.
Thank you!
You can also give us your feedback online. For further
information, visit:
www.dspace.com/magazine
Release information on dSPACE products is available at:
www.dspace.com/releases
dSPACE Magazine 1/2013 dSPACE GmbH, Paderborn, Germany info@dspace.com www.dspace.com
d
S
P
A
C
E
M
a
g
a
z
i
n
e
1
/
2
0
1
3
www.dspace.com
1/2013
MAGAZINE
dSPACE
Siemens Development Process
According ISO 26262
Mazda More Compression,
Less Fuel
Thales Alenia Space
Computers in Orbit
Get a Grip on Your Data
with dSPACE SYNECT